Not 2019 predictions, but what marketing AI developments do I want to see in 2019? This week we become selfish and ask for a lot.
Instead of predictions
I was thinking about what do I want to see happen in 2019 that could advance the use of AI in marketing? I don’t have any knowledge if any of the below will happen, but if they did happen in 2019, I would be very happy. Let’s just jump right in.
Salesforce introduces multi touch attribution
Multi touch attribution is still a mixed area. I was surprised to learn in surveys that adoption of multi touch attribution tends to be the same in B2C and B2B; both have about 40% adoption by marketers (if you are an eMarketer subscriber, there is a good link here), which based on experience, seems pretty high. I would guess many people are aware that some platform they are using, Google, Facebook, etc, does some sort of multi touch attribution, so they say, “Yes, I am doing it.” But, if you were given an incremental $100k to spend on a marketing campaign, could you confidently say you would know where to spend it?
I would like to see Salesforce take a crack at this problem. Salesforce is aggressively expanding into the marketing technology space and recently acquired a company called Datorama, for a cool $800mn. Datorama is a company that helps marketers pull all of their data together in one place, including everything from programmatic to Facebook data. It would be only logical that Salesforce could use this data from Datorama, combine it with their actual salesperson interaction and sales data in Salesforce, to build one really super cool attribution model.
Generative adversarial networks, or types of algorithms that generate an output, have become more realistic in their generations. When GANs were in their infancy, images created years ago used to be a little wonky, now they look lifelike.
Example of GAN created images
I would like to see an ad platform experiment with GANs to change advertisements.
I don’t expect these GANs to generate entirely new advertisements, but it would
be nice once you create an ad to be able to change certain elements of the ad.
Need a new background? Need a new color? GANs should be able to automatically
generate these things for you. The technology exists, now it’s just time to execute.
It’s a shame that we only hear about Marcel every six months. The last mention was probably not the one Publicis envisioned when it started prototyping Marcel: a “fake” holiday card that features people saying nasty thinks to Marcel. Each year Publicis does a spoof holiday card (Last year it featured the new CEO aging so much because of the stress) and there is always some truth behind the jokes.
I would like to see any sort of update, from how many people are using it, to are any clients using it; any sort of stat around adoption of this product would nice. If it’s terrible, then tell us. Despite all the negative attention, including some from Cloudy, Marcel could be the project that helps advertisers think more about AI and how to integrate it. It’s important that Marcel doesn’t die.
More automated ad cases
I would like to see more companies or agencies adopt automated ad buying. I don’t have a particularly good way of tracking adoption of this, as surveys are hard to come by, but the company AlbertAI, who got a ton of press off their Coabella survey, only had $2mn in revenue for the first half of 2018. Maybe people are uncomfortable with automating this aspect, or think the service is too expensive or too complicated, but the results are astoundingly good.
If you are concerned about how complicated the technology is, the problem is the technology is going to get even more complicated. Alibaba is experimenting with using reinforcement learning to bid for ads, which means you must be an expert in AI to understand how this all works. I also don’t exactly see the price coming down as you need world class talent to develop these sort of algorithms, so just take the plunge and automate.
The voice market to develop
I don’t think voice will become the most important advertising channel, but I would like to see some guidance from Amazon/Google about how their ranking works and if they are going to be implementing voice engine optimization (VEO)…which apparently the term has been trademarked. Being the first result that Alexa returns will probably be life or death, as I would not imagine many people aren’t interested in hearing more than 3 suggestions from Alexa.
I am also very interested in seeing how Amazon ranks its own products via Alexa. For example, if you search for “Coffee” on Amazon, the first row is a sponsored header for a single company, the second row is sponsored products from numerous companies and the third row is Amazon’s own white label coffee. Which means you don’t get any organic search terms until the fourth row. It will be interesting to see if Amazon takes the same approach with Alexa.
People to stop including “personalization” in their marketing trends predictions
I would like to see people stop including “personalization” in their marketing trend articles for 2019. These same articles will probably include personalization in their articles for 2020. The fact is, f you haven’t implemented personalization yet you are behind; the tools exist, the data is there, the case studies exist, just start doing it.
That’s all I want